Re: So, is COUNT(*) fast now? - Mailing list pgsql-hackers

From Tom Lane
Subject Re: So, is COUNT(*) fast now?
Date
Msg-id 27592.1319827734@sss.pgh.pa.us
Whole thread Raw
In response to Re: So, is COUNT(*) fast now?  (Robert Haas <robertmhaas@gmail.com>)
Responses Re: So, is COUNT(*) fast now?
List pgsql-hackers
Robert Haas <robertmhaas@gmail.com> writes:
> I also tried changing the BufferIsValid() tests in
> visibilitymap_test() to use BufferIsInvalid() instead, with the sense
> of the tests reversed (see attached vismap-test-invalid.patch).  Since
> BufferIsInvalid() just checks for InvalidBuffer instead of also doing
> the sanity checks, it's significantly cheaper.  This also reduced the
> time to about 330 ms, so seems clearly worth doing.

Hmm.  I wonder whether it wouldn't be better to get rid of the range
checks in BufferIsValid, or better convert them into Asserts.  It seems
less than intuitive that BufferIsValid and BufferIsInvalid aren't simple
inverses.
        regards, tom lane


pgsql-hackers by date:

Previous
From: Robert Haas
Date:
Subject: Re: Include commit identifier in version() function
Next
From: Robert Haas
Date:
Subject: Re: So, is COUNT(*) fast now?